Knowledge graph is a powerful tool to organize and analyze complex biological interactions in cancer research from unstructured text like cancer literature. This study provides a framework to build an ontology-driven knowledge graph from cancer literature using a hybrid approach by combining LLM-based triplet extraction with Gene Ontology-based triplets, ensuring semantic consistency. The constructed knowledge graph captures diverse molecular processes and signaling pathways related to the ’Sustaining proliferative signaling’ cancer hallmark. The study applied graph-theoretical analyses such as degree centrality, betweenness centrality, eigenvector centrality, PageRank, and HITS algorithms to identify important nodes and relationships in the graph network. The centrality measure analysis of the knowledge graph captures interesting patterns, such as key processes like apoptosis, cell cycle regulation, EGFR signaling, and p21-mediated control, aligning with the hallmark of sustaining proliferative signaling. Furthermore, the PageRank and HITS authority scores identified highly connected regulatory processes. This thorough graph analysis reveals previously undiscovered pathways of possible therapeutic interest in addition to confirming established carcinogenic drivers. Finally, identified biological concepts can serve as crucial features for machine learning and deep learning algorithms for performing various tasks on cancer literature, such as document classification, node prediction, and functional annotation. The proposed work demonstrates the value of knowledge graph construction and analysis in understanding cancer biology and guiding precision medicine strategies.